From: Jes Sorensen <Jes.Sorensen@redhat.com>
To: Shaohua Li <shli@kernel.org>
Cc: Artur Paszkiewicz <artur.paszkiewicz@intel.com>,
NeilBrown <neilb@suse.com>,
linux-raid@vger.kernel.org
Subject: Re: [PATCH v2 00/12] Partial Parity Log for MD RAID 5
Date: Tue, 03 Jan 2017 10:42:05 -0500 [thread overview]
Message-ID: <wrfjzij8ywtu.fsf@redhat.com> (raw)
In-Reply-To: <20161216232421.x6sb4goclzyasu6k@kernel.org> (Shaohua Li's message of "Fri, 16 Dec 2016 15:24:21 -0800")
Shaohua Li <shli@kernel.org> writes:
> On Thu, Dec 15, 2016 at 12:44:57PM +0100, Artur Paszkiewicz wrote:
>> On 12/14/2016 08:47 PM, Shaohua Li wrote:
>> > For the implementation, I don't understand how the ppl works much,
>> > there aren't
>> > many details there. Two things I noted:
>> >
>> > - The code skips the log for full stripe write. This isn't
>> > good. It would means
>> > after a unclean shutdown/recovery, one disk has arbitrary data,
>> > not the old
>> > data and new data. This breaks an assumption in filesystem, after a failed
>> > write to a sector, the sector has either old or new data. Thinking about a
>> > write to superblock. The data could be old or new superblock,
>> > but it's still a
>> > superblock, not something random.
>> >
>> > - From the patch 6 & 10, looks PPL only help recover unwritten disks. If one
>> > disk of a stripe is dirty (eg it's written before unclean
>> > shutdown), and it's
>> > lost in recovery, what will happen? Seems the data of lost disk
>> > will be read as
>> > 0? It will break the assumption above too. If I understand the
>> > code clearly
>> > (maybe not, need clarification), this is a design flaw.
>>
>> PPL is only used to update the parity for a stripe, data chunks are not
>> modified at all during PPL recovery. The assumption was that it would
>> protect only from silent data corruption, to eliminate the cases when
>> data that was not touched by a write request could change. So if a dirty
>> disk is lost, no recovery is performed for this stripe (parity is not
>> updated). For full stripe write we only recalculate the parity after a
>> dirty shutdown if all disks are available (like resync). So you are
>> right that it is still possible to have arbitrary data in the written
>> part of a stripe if that disk is lost. In such case the behavior is the
>> same as in plain raid5.
>
> Ok, this matches my understanding. This isn't a completed solution but does
> help a lot. If users want to use this, there is no reason to not support it.
> After you fix the alignment issue and describe the solution in details, I'll
> look at it again.
Artur,
Did you make any progress getting the alignment issue resolved?
I'd really like to get an mdadm release out the door this week, so
getting this resolved would be awesome. Hint hint ;)
Cheers,
Jes
next prev parent reply other threads:[~2017-01-03 15:42 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-05 15:31 [PATCH v2 00/12] Partial Parity Log for MD RAID 5 Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 01/12] raid5-cache: move declarations to separate header Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 02/12] raid5-cache: add policy logic Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 03/12] raid5-cache: add a new policy Artur Paszkiewicz
2016-12-07 0:46 ` NeilBrown
2016-12-07 14:36 ` Artur Paszkiewicz
2016-12-07 23:24 ` NeilBrown
2016-12-08 10:28 ` Artur Paszkiewicz
2016-12-08 21:22 ` NeilBrown
2016-12-05 15:31 ` [PATCH v2 04/12] md: superblock changes for PPL Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 05/12] raid5-ppl: Partial Parity Log implementation Artur Paszkiewicz
2016-12-06 1:06 ` kbuild test robot
2016-12-07 1:17 ` NeilBrown
2016-12-07 14:37 ` Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 06/12] raid5-ppl: calculate partial parity Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 07/12] md: mddev_find_container helper function Artur Paszkiewicz
2016-12-07 1:23 ` NeilBrown
2016-12-05 15:31 ` [PATCH v2 08/12] md: expose rdev->sb_start as sysfs attribute Artur Paszkiewicz
2016-12-07 1:25 ` NeilBrown
2016-12-05 15:31 ` [PATCH v2 09/12] raid5-ppl: read PPL signature from IMSM metadata Artur Paszkiewicz
2016-12-07 1:25 ` NeilBrown
2016-12-07 14:38 ` Artur Paszkiewicz
2016-12-07 23:27 ` NeilBrown
2016-12-08 10:36 ` Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 10/12] raid5-ppl: recovery from dirty shutdown using PPL Artur Paszkiewicz
2016-12-05 15:31 ` [PATCH v2 11/12] raid5-ppl: support disk add/remove with distributed PPL Artur Paszkiewicz
2016-12-07 1:29 ` NeilBrown
2016-12-05 15:31 ` [PATCH v2 12/12] raid5-ppl: runtime PPL enabling or disabling Artur Paszkiewicz
2016-12-07 0:32 ` [PATCH v2 00/12] Partial Parity Log for MD RAID 5 NeilBrown
2016-12-07 14:36 ` Artur Paszkiewicz
2016-12-07 17:09 ` Shaohua Li
2016-12-13 15:25 ` Jes Sorensen
2016-12-14 19:47 ` Shaohua Li
2016-12-15 11:44 ` Artur Paszkiewicz
2016-12-16 23:24 ` Shaohua Li
2017-01-03 15:42 ` Jes Sorensen [this message]
2017-01-04 8:01 ` Artur Paszkiewicz
2017-01-04 13:29 ` Jes Sorensen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=wrfjzij8ywtu.fsf@redhat.com \
--to=jes.sorensen@redhat.com \
--cc=artur.paszkiewicz@intel.com \
--cc=linux-raid@vger.kernel.org \
--cc=neilb@suse.com \
--cc=shli@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).